Abatacept and reduced immune response to pandemic 2009 influenza A/H1N1 vaccination in patients with rheumatoid arthritis


  • ClinicalTrials.gov identifier: NCT01151644.



To evaluate the influence of abatacept (ABA) and associated contributing factors on pandemic 2009 influenza A/H1N1 vaccine immunogenicity in rheumatoid arthritis (RA) patients.


The response to a nonadjuvanted monovalent pandemic 2009 influenza A/H1N1 killed virus vaccine was analyzed in 11 RA patients using ABA (RA-ABA), most with concomitant nonbiologic disease-modifying antirheumatic drugs (DMARDS), and compared to 33 age-matched RA patients on methotrexate (MTX) and 55 healthy controls, all without previous seroprotection. Clinical and laboratory evaluations were performed before and 21 days after vaccination. Anti-influenza antibody titers were measured by hemagglutination inhibition assay. Seroprotection (antibody titers ≥1:40) and the factor increase (FI) in the geometric mean titers (GMTs) were calculated. Prevaccination lymphocyte counts and gammaglobulin levels were determined.


Sex distribution, disease duration, and the Disease Activity Score in 28 joints were similar in the RA groups (P > 0.05). After vaccination, seroprotection was significantly reduced in RA-ABA patients compared to RA-MTX patients (9% versus 58%; P = 0.006) and controls (69%; P ≤ 0.001). FI-GMT was severely reduced in RA-ABA patients compared to RA-MTX patients (1.8 [1.4–2.3] versus 8.7 [5.2–17.4]; P < 0.001) and controls (11.5 [8.0–16.7]; P ≤ 0.001). Lymphocyte counts were comparable in RA groups (P > 0.05), but RA-ABA patients had slightly lower gammaglobulin levels than RA-MTX patients (0.9 gm/dl [0.6–1.8] versus 1.2 gm/dl [0.8–1.7]; P = 0.03), although almost all were within the normal range values.


The current study established that ABA, in association with traditional DMARDs, significantly reduces the humoral response to pandemic 2009 influenza A/H1N1 vaccine in RA patients. The results suggest an influence of costimulatory modulation in humoral response to this vaccine.


Rheumatoid arthritis (RA) treatment can induce immunosuppression and increase the risk of infection (1). Abatacept (ABA) is a soluble fusion protein that selectively modulates the CD80/CD86:CD28 costimulatory signal required for full T cell activation (2).

Recently, our group studied the response to a nonadjuvanted antipandemic 2009 influenza A/H1N1 vaccine in a large Brazilian cohort of 1,668 autoimmune rheumatic disease patients. The results demonstrated a diminished immunogenicity to this vaccine in RA patients, especially those undergoing methotrexate (MTX) therapy (3, 4). Other trials involving the use of adjuvanted pandemic influenza vaccines also showed that the use of disease-modifying antirheumatic drugs (DMARDs) (5), including MTX (5, 6) and leflunomide (5, 7), could impair the response. In regard to biologic therapies, tumor necrosis factor inhibitors were reported to have a limited impact on the pandemic influenza vaccination (4–7), while rituximab was associated with a significant reduction on the humoral response (5, 6). There are, however, limited data about the impact of ABA in the influenza vaccine with only 1 report using adjuvanted vaccine, suggesting a negative effect (6). The study included a heterogeneous group of patients, which may have hampered the interpretation of findings because diverse immunogenicity related to rheumatic disease was observed in patients immunized with the pandemic vaccine (3, 6, 7). Moreover, the inclusion of a specific age-matched healthy control group and an RA group under treatment only with traditional DMARDs is essential to minimize the influence of age (8) and discriminate the effect of each therapy (4–7).

Thus, the aim of this report is to assess the influence of ABA and associated contributing factors on the immune response to a nonadjuvanted pandemic 2009 influenza A/H1N1 vaccine. The results were compared to those of age-matched healthy controls and RA patients using only traditional DMARDS.

Significance & Innovations

  • The efficacy of pandemic influenza A/H1N1 vaccines has not been properly evaluated in rheumatoid arthritis (RA) patients under treatment with abatacept.

  • The immunogenicity of a nonadjuvanted pandemic 2009 influenza A/H1N1 vaccine in RA patients treated with such biologic therapy was studied and compared to age-matched RA patients treated with methotrexate and to healthy controls.

  • Prevaccination lymphocyte counts and gammaglobulin levels were also evaluated.

Patients and methods

Study design and participants.

This is a subanalysis of a prospective study conducted at a single center in São Paulo, Brazil, during the Public National Health pandemic 2009 influenza A/H1N1 vaccination campaign, as described elsewhere (3). It was approved by the Institutional Review Board. In the prior study, 340 RA patients (according to the 1987 American College of Rheumatology criteria) (9), ages ≥18 years and seen for regular followup at the Rheumatology Outpatient Clinic of the Hospital das Clínicas da Universidade de São Paulo, were included and compared to healthy controls. In this subanalysis, 11 RA patients treated with ABA (RA-ABA) were compared to age-matched control groups, i.e., RA patients treated with MTX (RA-MTX) and healthy controls. The exclusion criteria were described elsewhere (3, 4).

The vaccine, Sanofi Pasteur Influenza A/H1N1, was a nonadjuvanted monovalent pandemic 2009 influenza A/H1N1 killed virus vaccine (A/California/7/2009/Butantan Institute/Sanofi Pasteur, São Paulo, Brazil) containing 15μg hemagglutinin from an influenza A/California/07/2009(H1N1) virus-like strain (NYMCx-179A) per 0.5-ml dose (3, 4).

Study procedures.

All subjects received a single intramuscular dose of the vaccine. Blood samples were collected before and 21 days after the vaccination; the erythrocyte sedimentation rate (ESR), C-reactive protein level, and antibody analyses, as well as the clinical evaluation of disease activity (Disease Activity Score in 28 joints using the ESR) were performed. Prevaccination collection of whole blood for lymphocyte count (automated standard cell blood count and differential) and for analysis of total immunoglobulin levels (semiquantitative electrophoretic separation by densitometry) was also performed. Pre-ABA lymphocyte counts and total immunoglobulin levels, measured by the same techniques, were available by routine screening for patients beginning/changing any biologic treatment and were retrospectively assessed. Regarding safety assessments, a 21-day symptom diary card for prospective completion was given to each participant following vaccination and returned 21 days after vaccination (3, 4).

Hemagglutination inhibition assay.

The influenza antigen used was the H1N1 A/California/7/2009 supplied by the Butantan Institute. Antibody titers were obtained at baseline and 21 days after immunization in all groups. Two serologic end points were calculated: the postvaccination seroprotection rate (percentage of patients with titers ≥1:40) and the seroconversion rate (percentage of patients with a ≥4-fold increase in vaccination titers if the prevaccination titers were ≥1:10 or titers ≥1:40 if the prevaccination titers were <1:10). The antibody geometric mean titers (GMTs) and the factor increase (FI) in the GMTs were also calculated.

Statistical analysis.

Age-matching of the RA-MTX group and healthy controls was carried out by random selection using SPSS software, version 15. GMTs and FI-GMT were calculated and analyzed using log-transformed data. Categorical variables were compared by Fisher's exact test or chi-square test as appropriate. Normally or non-normally distributed variables were compared using the t-test and Wilcoxon's rank sum test, respectively. When comparisons of continuous variables were performed among more than 2 groups, one-way analysis of variance or Kruskal-Wallis analysis of variance was used as appropriate. All tests were 2-sided (alpha level of 0.05).


Baseline data.

The 11 RA-ABA patients were age-matched to random RA-MTX patients (n = 33) and healthy controls (n = 55), and all of them completed the study. To overcome previous seroprotection as a confounder, none of the subjects included had discernible levels of seroprotective antibodies before vaccination. RA-ABA patients were similar to RA-MTX patients regarding age, sex, disease duration, disease activity, inflammatory markers, concomitant use of leflunomide or chloroquine, and total number of DMARDs taken. There was, however, a lower frequency of MTX use in patients under treatment with ABA (P < 0.05). The healthy control group had comparable age and female sex predominance (Table 1).

Table 1. Demographic characteristics and disease activity parameters before pandemic 2009 influenza A/H1N1 vaccination and therapy of RA patients at baseline*
 RA-ABA (n = 11)RA-MTX (n = 33)Controls (n = 55)P
  • *

    Values are the absolute number (frequency) or the median (range). RA = rheumatoid arthritis; ABA = abatacept; MTX = methotrexate; RF = rheumatoid factor; anti-CCP = anti–cyclic citrullinated peptide; DAS28 = Disease Activity Score in 28 joints; ESR = erythrocyte sedimentation rate; CRP = C-reactive protein; GC = glucocorticoids; LEF = leflunomide; CHLOR = chloroquine; DMARDs = disease-modifying antirheumatic drugs.

  • Related to concomitant comparison of the 3 groups.

Female11 (100)29 (88)42 (76)0.11
Age, years55 (27–75)56 (29–74)52 (27–80)0.49
RA parameters    
 RF positivity5 (45)25 (76)0.13
 Anti-CCP positivity6 (55)28 (85)0.26
 Disease duration, years17 (8–28)12 (1–34)0.19
 DAS28-ESR3.8 (2.5–6.0)3.6 (1.2–7.2)0.98
 ESR, mm/hour34 (7–78)40 (0–100)0.06
 CRP level, mg/dl5.9 (1.5–41.7)7.8 (1.0–53.4)0.67
 GC9 (82)28 (85)1
 GC dose, mg/day7.5 (5–10)10 (2.5–40)0.89
 MTX6 (55)33 (100)< 0.001
 MTX dose, mg/week25 (20–25)25 (15–25)0.43
 LEF4 (36)6 (18)0.24
 CHLOR3 (27)15 (46)0.48
 DMARDs2 (1–3)2 (1–3)0.11


RA-ABA patients received the recommended doses according to weight (2). The median (range) duration of treatment was 34 weeks (4–273 weeks). Eight (73%) patients had been taking ABA for more than 24 weeks. The median (range) time since the last dose was 18 days (0–21 days) prior to the vaccination. Nevertheless, as ABA is administered every 4 weeks, each patient received 1 dose during the 3-week period of the study.

Response to immunization.

Prevaccination GMTs were very low and similar in all groups (Table 2). Seroconversion was not obtained in any of the RA-ABA patients, and only 1 subject (9%) achieved seroprotection. These trends were significantly different from those observed in other groups (P < 0.001 for seroconversion and P = 0.001 for seroprotection). Despite a significant and slight increase in GMT (6.0 [95% confidence interval (95% CI) 4.6–7.9] to 10.7 [95% CI 7.2–15.7]; P = 0.008) after vaccination, FI-GMT (P < 0.001) and postvaccination GMT (P < 0.001) were severely reduced in the RA-ABA group compared to the other groups (Table 2). RA-MTX patients and controls had more significant increases in GMT after vaccination (6.0 [95% CI 5.3–6.9] to 52.6 [95% CI 31.5–87.7]; P < 0.001 and 6.6 [95% CI 5.8–7.5] to 76.1 [95% CI 52.9–109.3]; P < 0.001, respectively). In all parameters analyzed, RA-MTX patients exhibited lower responses than controls, but these differences did not reach statistical significance. No correlation was observed between any of the end points and the duration of treatment with ABA or time since the last dose (P > 0.05).

Table 2. Serologic data before and after pandemic 2009 influenza A/H1N1 vaccination of RA patients and controls*
 RA-ABA (n = 11)RA-MTX (n = 33)Controls (n = 55)P
  • *

    Values are the absolute number (frequency) unless otherwise indicated. RA = rheumatoid arthritis; ABA = abatacept; MTX = methotrexate; GMT = geometric mean titer; 95% CI = 95% confidence interval; FI = factor increase in GMT after vaccination. GMT significantly increased in all RA subsets and in controls (P < 0.05).

  • Related to the concomitant comparison of all 3 groups.

  • P < 0.001 (RA-ABA vs. RA-MTX); P ≤ 0.001 (RA-ABA vs. controls).

  • §

    P = 0.006 (RA-ABA vs. RA-MTX); P ≤ 0.001 (RA-ABA vs. controls).

Before vaccination    
 GMT (95% CI)6.0 (4.6–7.9)6.0 (5.3–6.9)6.6 (5.8–7.5)0.77
After vaccination    
 Seroconversion019 (58)36 (66)< 0.001
 Seroprotection1 (9)§19 (58)38 (69)0.001
 GMT (95% CI)10.7 (7.2–15.7)52.6 (31.5–87.7)76.1 (52.9–109.3)< 0.001
 FI-GMT (95% CI)1.8 (1.4–2.3)8.7 (5.2–17.4)11.5 (8.0–16.7)< 0.001

Lymphocyte counts and gammaglobulin levels.

Prevaccination lymphocyte counts were similar in both RA groups (P = 0.73) (Table 3). In contrast, RA-ABA patients had prevaccination gammaglobulin levels lower than those observed in RA-MTX patients (P = 0.03) (Table 3), but only 1 patient presented with a value (0.6 gm/dl) below normal range (0.7–1.5) and 4 patients presented with a borderline level (0.7 gm/dl), a value at the limit of normal range. The RA-MTX group had no patient with low or borderline gammaglobulin levels. Of note, the 8 patients under treatment with ABA for more than 24 weeks had significantly lower median gammaglobulin levels compared to those treated for a shorter length of time (0.7 gm/dl [range 0.6–1.2] versus 1.1 gm/dl [range 1.1–1.8]; P = 0.048). In addition, gammaglobulin levels measured and available in all RA-ABA patients immediately before ABA treatment were within the normal range and statistically higher (P = 0.004) than prevaccination gammaglobulin levels, with no patient showing low or borderline levels immediately before ABA treatment (Table 3).

Table 3. Lymphocyte counts and gammaglobulin levels before vaccination in the RA groups and pretreatment gammaglobulin level in the RA-ABA group*
 RA-ABA (n = 11)RA-MTX (n = 33)
  • *

    Values are the median (range). RA = rheumatoid arthritis; ABA = abatacept; MTX = methotrexate.

  • P = 0.03, RA-ABA vs. RA-MTX; P = 0.004, prevaccination vs. pre-ABA levels within the RA-ABA group.

Prevaccination lymphocyte count (cells/mm3) (normal range 900–3,400)2,000 (1,000–4,000)2,100 (800–5,000)
Prevaccination gammaglobulin level (gm/dl) (normal range 0.7–1.5)0.9 (0.6–1.8)1.2 (0.8–1.7)
Pre-ABA gammaglobulin level (gm/dl) (normal range 0.7–1.5)1.3 (0.9–2.3)

Regarding adverse events, severe side effects were not reported during the followup period. The rates of minor side effects were comparable: 55% in RA-ABA patients, 39% in RA-MTX patients, and 40% in control groups (P = 0.64).


The current study established that ABA significantly reduces the humoral response to pandemic 2009 influenza A/H1N1 vaccine in RA patients. The slightly lower immunoglobulin levels observed in these patients could partially explain this alteration.

The remarkable impact of costimulatory modulation on the humoral response was evidenced by all the parameters analyzed herein in RA-ABA patients: the uniform lack of seroconversion, the striking reduction in seroprotection, and the low GMTs and FI-GMT. These findings reinforce a single previous published study using adjuvanted pandemic 2009 influenza A/H1N1 vaccine, which also demonstrated a reduced humoral response in a heterogeneous group of inflammatory arthritis patients under ABA treatment (6). In fact, studies with animal models revealed a critical role for CD80/86:CD28 costimulation during the immune response to different antigens (10, 11), and the response to the T cell–dependent neoantigens bacteriophage X174 and keyhole limpet hemocyanin (KLH) was reduced in psoriasis patients treated with ABA, without inducing tolerance (12).

In contrast, few trials failed to show a major reduction in the immune response to other antigens. Healthy individuals who received a single dose of ABA had a decreased, but not significantly inhibited, response to tetanus toxoid and 23-valent pneumococcal vaccines (13), probably explained by the very limited ABA exposure, without combination therapy. Furthermore, studies with pneumococcal (14) and seasonal influenza (15) vaccines with a limited number of RA patients were performed, without control groups, suggesting an adequate response. Regarding pneumococcal vaccine, the polysaccharide and less T cell–dependent nature of the antigen may account for the preserved immune response during costimulatory modulation with ABA. Regarding seasonal influenza vaccine, previous exposure to such viral or vaccine antigens may explain the results, whereas none of the subjects evaluated herein were seroprotected before immunization, which is what makes this pandemic influenza vaccine comparable to a pure neoantigen. It remains to be determined if a second boost would improve vaccine response in these patients as demonstrated previously in patients under DMARDs and other biologic therapies (5).

This result is strengthened by the inclusion of the relevant control groups since RA patients frequently use ABA in combination with traditional DMARDs. Because MTX was associated with a reduced response to the pandemic influenza vaccine (4–6), it was therefore crucial to include an RA age-matched group under treatment with MTX for comparison. Additionally, an age-matched healthy control group was necessary to avoid the known influence of this parameter on pandemic influenza vaccine immunogenicity (8).

Lymphopenia was not detected in these patients and cannot account for the compromised vaccine immunogenicity, whereas the observed slightly lower gammaglobulin levels in RA-ABA patients may have influenced pandemic influenza vaccine response or reflected the same subjacent mechanism. Responses to the pandemic influenza vaccine and to the neoantigen KLH were also reported to be decreased in RA patients treated with rituximab, a well-established inhibitor of the humoral immune system (5, 6, 16).

An indication bias favoring a correlation between ABA use and reduced gammaglobulin levels in patients referred for ABA therapy does not seem to be a likely explanation for the lower gammaglobulin levels in RA-ABA patients, given that the pre-ABA gammaglobulin levels were within the normal range, with no patient showing low or borderline levels. On the other hand, the significant prevaccination lower levels of gammaglobulin, although within normal range for all but 1 patient, seemed related to long-term ABA use. These novel findings need to be confirmed as a drug effect. In addition, we cannot exclude the drug synergism between ABA and traditional DMARDs, since the majority of patients were under combination therapy.

The small number of patients represents a limitation of the findings. Convenience samples are prone to selection bias and low extern validation. On the other hand, such a severe reduction in humoral responses in a small sample may imply a robust drug effect.

In conclusion, inhibition of costimulation mediated through CD80/86:CD28 in association with traditional DMARDs was related to a severe reduction in the humoral response to influenza A/H1N1 vaccine in RA patients.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Ribeiro had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Ribeiro, Laurindo, Guedes, Saad, Moraes, Silva, Bonfa.

Acquisition of data. Ribeiro, Laurindo, Guedes, Saad, Moraes, Silva, Bonfa.

Analysis and interpretation of data. Ribeiro, Laurindo, Bonfa.